How to Sync Audio and HDV video

[Ben Moore]

Originally Posted by Nelson Alexandre

That's exactly right Ben. And for what it matters, I think that Tascam's recorder is THE bang for the buck concerning flash based recorders. So why do you need any others? ;-)

Thats a good point, I have heard a lot of good things about it. I just kinda wish the Edirol R-4 had it. Its 4 channels and the one I have been looking at.

For what its worth The Marantz PMD620 keeps great sync with video, when using 48khz WAV files. I have kept sync up to an hour many times.

Ben

[Mike Demmers]

Originally Posted by Don Xaliman

It's got me thinking.
I have not noticed much sync drift in my concert recordings but a bit does appear on the +15 min compositions. So in my quest to better use the equipment I have, it struck me that since I often record separate audio on a Tascam DA-38 8 track digital tape recorder that the simple act of hooking a cable between my Canon XH-A1 composite out and the word sync in, on my Tascam deck, I will accomplish much better sync. Assuming that the length won't vary by the time I capture the audio into the computer, mix it in Wavelab and place the mixed wave file into my video edit timeline.
At least it will be an improvement over the machines all going on their own clock.
Thanks for the insight.

Ah, I hope you read the rest of this thread and understand you cannot just plug a video output into a word clock input wthout a converter in between.

I think these had a video sync card option though (blackburst), do you have that?

-Mike

[Don Xaliman]

I think these had a video sync card option though (blackburst), do you have that?

-Mike
----------------------
That option is not available on my model, the DA-88 and 98 model had that option.
I re-read the post in this topic and am now clear that if I ever need right on sync, then more gear can be acquired. As for now, editing the video and synchronizing the audio from the post mix, one composition at a time has been fairly forgiving and the drift very small, much smaller than the - "within a half of a frame" - offset that often appears in lining the clips up on the frame based timeline of NLE's.
... is the audience really listening ???.

[Mario Vermunt]

I just found two cheap Video Slave Drivers from Protools (these convert blackburst from the composite output of the camera to wordclock that can be fed to my SD702 recorder).

They are rackmount 220V devices, but I will try to convert them in a new housing with battery feed. If, or rather when ;)) I succeed I will have one spare for another dvinfo user.

[Jay Massengill]

Good find Mario!
Of course the US version was 120v.
Another note, they could do 44.1k or 48k sample rate, but nothing higher than that.

[Mike Demmers]

Originally Posted by Mario Vermunt

I just found two cheap Video Slave Drivers from Protools (these convert blackburst from the composite output of the camera to wordclock that can be fed to my SD702 recorder).

They are rackmount 220V devices, but I will try to convert them in a new housing with battery feed. If, or rather when ;)) I succeed I will have one spare for another dvinfo user.

Great find!

These should be really simple circuits. If you are lucky the actual PC card may be quite small and could be moved to a smaller box.

-MD

[Ray Drueke]

After several Google searches around this sync topic I've managed to perhaps find a production based solution but can't find the corresponding post-production (editing) solution.

Help?

Background:
I'm working to record live music performances with up to 3 miniDV cameras for video and up to 16 tracks of audio on a hard disk recorder. The audio is brought to the hard disk recorder via mic splitters that give me a separate feed for each mic or input on stage. I don't want/can't connect the cameras with cable and don't have access to live switching equipment. (I'd want to record all the audio and video as a backup anyway.)

What I've found is SMPTE LTC - an audio space time code that can be recorded to an unused track on the hard disk recorder and recorded on an audio track on each camera. Using something like an MOTU Timepiece and wireless transmission would deliver the same time code on audio tracks on each recording device - the hard disk recorder and the cameras.

What I haven't found (yet) is how to sync these 'feeds' in post production editing. I have ProTools, Audition and Premiere Elements available. I can bring in (capture) each 'feed' but how do I tell Premiere that this audio (left or right) is time code? I could, if necessary, bring the (real) audio tracks into ProTools, mix and then create 2 new 'feeds' for left + LTC and right + LTC.

Is there any hope for my approach? If so I'd appreciate any ideas about the syncing end of all this. If I can get these 'feeds' sync'd then I can edit to the client's desire.

Thanks,
Ray.

[Seth Bloombaum]

I've not read back in this older thread, but do lots of this, both for my productions and for others'. This last weekend I was recording 24-track, 11 bands, 3 cameras.

I don't use external TC on any device. I use camcorders' internal TC generators to set time-of-day (free-run) timecode as closely as possible - this is typically within 10 frames.

If the audio recorder supports timecode via BWF-stamping, I use it in the same way as above. If it doesn't, (eg., my 24-track doesn't), I don't worry about it.

I don't use MOTU or any other synchronizer in my workflow. Once all these video and audio clips are on the timeline the synchronization takes place right there. My post environment is Vegas, which is very, very good for this type of work. Having not used Premiere for 10 yrs. or more, I don't know how well the following applies. Note that PluralEyes sync software has just been beta released for Vegas, up till now it's been FCP.

My workflow works best with a big harddrive in the audio recorder, and run the recorder continuously from start to finish. Depending, you may run into a tricky 4GB size limitation on wav files. Utilities are needed to split such files on ingest - I use the fantastic HD24tools for the Alesis HD24. If there's no way to split your files, you'll need to make sure none of them go over 4GB, period. Another safety measure - record a mix on a separate recorder.

0) Record reference audio on all cameras - very important!
1) Video clips on timeline.
2) Turn on TC thumbnail display on all clips.
3) Set the timeline's start time to some convenient time a few minute before your first video clip's start.
4) Using timeline navigation tools, go to the precise time that matches each clip's starttime. Drag the clip and snap it to the cursor. Rough video sync - done!
5) If audio is time-stamped BWF, you may have a file import function that parks it on the timeline at the right spot.
6) If no TC stamp on audio, look at wave-forms for song starts on your camera reference audio vs. 2nd system audio, and drag the audio-only recording to match. Rough audio sync - done!
7) Fine sync - you're always moving the audio-only clip(s). Listen to a mix of them against one camera's sound, and slip to reduce and eliminate echo. Fine sync - done!
8) Lock it, group it, save it, whatever. Keep this version of the project for reference, and move on to your multicamera edit.

Vegas has great tools supporting this sort of workflow. I hear that there are third-party tools for FCP that support audio-track-to-timecode-display. PluralEyes does syncing via waveform matching.

Premiere? You'll have to find a workflow, maybe it will do what I've described above...

PS. Longer takes are always better. Burning through a 1 hour tape is better than having to do 12 syncs because the operator started/stopped the cam.

PPS. Always record in 48KHz sample rate for best sync results.

PPPS. If your recorder's clock doesn't closely match your cameras, you'll have some problems with long takes. Depending on how Premiere can stretch/shrink the audio-only clip this may be incredibly easy or impossible to resolve. Best to figure this out before the shoot, do a test recording for an hour with a clap at the start and finish and see what the timebase error is at the end. Possible solutions: Resolve in Premiere. Use an audio recorder that will slave to an external clock, and convert a camera's black burst to whatever the audio recorder will understand. Use Vegas, at least for the syncing part of the project.

[Nicholas de Kock]

Plural Eyes will sync your audio/video for you automatically...
Singular Software

[Ray Drueke]

Seth,
Thanks a LOT for your quick reply - I appreciate it! I have a couple questions.

What's "BWF stamping?" I'm not familiar with that one and probably lots of others as well but...

I have an Alesis HD24 with 2 360GB drives and will look into "HD24tools." I mix thru a Mackie 1604 and have been putting a rough mix on my Canon XH/A1 which is usually physically next to the Mackie/Alesis or at least within a cable or two distance. The other 2 camcorders are old Sony consumer models and currently capture audio via internal mics.
I've used a single clap on stage in front of a vocal mic with all cameras in view in the past and this is barely acceptable from an artist's perspective. I don't have a slate but would use one if that's more acceptable. Once per 60 min. tape wouldn't be too bad.

Thanks again for your ideas/suggestions/workflow - I appreciate it.

Ray.

[Steve House]

Originally Posted by Ray Drueke

...I'm working to record live music performances with up to 3 miniDV cameras for video and up to 16 tracks of audio on a hard disk recorder. The audio is brought to the hard disk recorder via mic splitters that give me a separate feed for each mic or input on stage.....

What cameras and recorder are you using? "Drift" over longer shots may or may not be a problem. Five minutes may work but an hour might not. What I'm wondering is whether the cameras you plan on using have inputs and outputs for their internal timecode and whether they support genlock inputs? Also, is the audio recorder file-based producting wav files or is a dat and whether it supports timecode natively? A common approach for your sort of scenario where sync must be maintained over a full concert is to use sync-boxes such as Ambient Lock-its that are tuned to each other providing timecode and sync to each camera and wordclock plus timecode to the audio recorder.

[Seth Bloombaum]

Originally Posted by Ray Drueke

Seth,
Thanks a LOT for your quick reply - I appreciate it! I have a couple questions.

What's "BWF stamping?" I'm not familiar with that one and probably lots of others as well but...

I have an Alesis HD24...

Recorders that write to Broadcast Wave Format (BWF) will usually provide an internal TC generator. The BWF specification includes a stamp in the header for starting timecode for each clip.

But this doesn't apply to your HD24. Which is a GREAT TOOL for this sort of recording, even though it doesn't support TC.

Run at 48KHz, the HD24 has been proven by many to have an extremely stable and standard clock. (not as good at 44.1, but you want 48 for video anyways)

Originally Posted by Ray Drueke

...The other 2 camcorders are old Sony consumer models and currently capture audio via internal mics.
I've used a single clap on stage in front of a vocal mic with all cameras in view in the past and this is barely acceptable from an artist's perspective...

OK, now I'm gonna' reveal just how long I've been working... a trick from olden tymes for dealing with non-TC, non-sync devices when filming concerts, like, a silent 16mm film camera, is to procure a large clock with a second hand and put it off to the side of the stage somewhere, where every camera can see it.

Now, each cam op zooms in on the clock whenever they roll tape, and catches a couple seconds. Voila, time-of-day code on a non-TC consumer camera at the head of every clip.

Another trick ancienne, the bloop-light. Get a friendly engineer to build you a box with a button, a small light bulb, and an oscillator. Press the button, the light goes on, the oscillator sends tone at line level. In operation, the camera op would come back to the wherever the recorder is (FOH?), and bloop with the lens pointed at the light. Tone goes to an unused HD24 channel. You might want to log each bloop using the HD24's h:m:s counter, so you know which camera blooped which tape.

Ah, yet another trick, not so ancient - start your cameras on 4-minute intervals, so, not everyone is changing tape at the same time.